Monitor datastore usage before it fills up.

When a VMware datastore fills up, ESXi pauses the VMs that want to write there. It usually shows in advance: snapshots grow, thin provisioning promises more than there is, a VM writes more every day. Spherifyer sees the trend, forecasts when the datastore will be full and warns in time with a recommendation.

Explore Spherifyer Documentation

01WarningsWhat Spherifyer checks on every datastore

Not just how full, but how fast.

A datastore at 85 % can have been stable for months or be full within a week. The difference lies in the history. Spherifyer measures on every collection run and calculates from the growth of the last 30 days when space runs out. The time to full is shown on every datastore card.

Usage: Warning from 75 %, critical from 90 %.
Forecast: Notice when the datastore will be full within 90 days at the current growth, warning from 30 days, with growth per day and free space.
Overcommit: Warning when the VMs have been promised more than the datastore holds (thin provisioning above 100 %).
Latency: Read or write from 5 ms elevated, from 20 ms critical, plus the VMs with the highest latency.
VMs per datastore: Notice from 16, warning from 21 active VMs on one datastore.
Veeam: Veeam replicas and backup snapshots are marked. Disks a Veeam proxy attaches via hot-add during a backup do not count as overcommitment.

02SnapshotsFind old snapshots

The forgotten snapshot is the most common space hog.

A snapshot before the update, then it is forgotten. Its delta file keeps growing with every change to the VM, takes up space and slows the VM down. Spherifyer lists every snapshot with name, age and creation date as a warning and shows snapshots older than 30 days in the reclamation list with the storage they use.

History: Usage over a selectable period from 30 days to more than 2 years: change per datastore, total used, provisioned and capacity.
Reclamation: Long powered-off VMs, idle VMs and old snapshots, with the total storage, vCPUs and RAM, optionally with an estimated saving.
Guest disks: Warning when a disk in the guest has less than 10 % free, critical below 5 %.

03FAQFrequently asked questions about datastore usage

What happens when a VMware datastore is full?

ESXi pauses VMs that want to write to the datastore until space is free again. This happens especially quickly with thin provisioning and growing snapshots. Spherifyer therefore warns from 75 % usage and forecasts when the datastore will be full.

How does Spherifyer calculate when a datastore will be full?

From the growth of the last 30 days, measured on every collection run. If the datastore will then be full within 90 days, a notice appears, from 30 days a warning. The 90-day limit is configurable.

How do I find old snapshots in vSphere?

Spherifyer reports every snapshot with name, age and creation date among the warnings. The reclamation list shows all snapshots older than 30 days with the storage they use, across all datacenters.

What does overcommitment mean for a datastore?

With thin provisioning a VM only uses what it actually writes, but it has been promised the full disk size. If the sum of these promises exceeds the capacity, the datastore is overcommitted. That is allowed, but risky when the VMs grow. Spherifyer warns from 100 % provisioning.

Do Veeam backups distort the values?

Without correction, yes: during a backup a Veeam proxy attaches the VM’s disks via hot-add, and vCenter counts them twice. Spherifyer subtracts this part and marks Veeam replicas and backup snapshots separately.